Model for Prediction of Saltwater Intrusions Based on Coupling of Support Vector Machine and Partial Least Square Method
Chen Xiao-hong · Acta Scientiarum Naturalium Universitatis Sunyatseni · 2007
The factors affecting the saltwater intrusions are analyzed by means of partial least square(PLM) method to extract the most important components,for solving multiple correlation among variables.Besides,the support machine vector(SVM) method is used for its advantage of solving non-linearity and little samples.A model for prediction of saltwater intrusions(PLS-SVM) based on the coupling of SVM and PLM is established and also successfully used in simulation and prediction of saltwater intrusion at Pinggang station in Zhuhai city.The results show that the PLS-SVM model is better than BP artificial neural network and Multiregression Models in the result of simulating and prediction.So it can be used for forecasting of saltwater intrusions.